Turning away the wrong customer on purpose

The catalog is professional dilutable concentrate: strong, unscented, sold in volumes a facility gets through in a month and a household would take nine years to finish. Roughly four in ten orders were coming from households who had found it through generic cleaning searches, bought it at a price that looked like a bargain per gallon, and were then genuinely unhappy — the smell, the strength, the size of the container, the fact that it must be diluted at all.

Those customers never reordered and almost always reviewed. The hero ASIN sat at 3.6 on a body of one- and two-star reviews that were, in their own terms, entirely fair. The rating then did the rest of the damage: facility buyers evaluating a chemical supplier read 3.6 as a quality signal and went elsewhere, so the wrong customer was actively driving away the right one.

We stopped trying to convert everyone and started filtering at the top.

• Deterrent copy by design — the first bullet and the second image state that this is a commercial concentrate requiring dilution, in the plainest possible terms.
• Search presence pulled back from domestic terms, with household cleaning keywords negated across the account rather than harvested.
• Volume and dilution shown as the scale they are — a jug beside a mop bucket, not beside a kitchen counter.
• Professional context everywhere — dilution ratios, contact times and surface use written for someone doing this as a job.

The hero ASIN's rating climbed from 3.6 to 4.0 over nine months as the reviews started coming from the people the product is built for.

Conversion rate fell 24%, exactly as intended — a large share of the clicks that used to convert should never have. Revenue still grew 34%, because facility buyers converted far more often once the rating and the copy told them this was a serious product, and because repeat orders went from a fifth of the business to nearly two thirds. A household buys once and complains; a facility buys every six weeks.